<<

Basic furnace based processes and cleanliness control at Baosteel

L. Zhang*1, J. Zhi2, F. Mei2, L. Zhu2, X. Jiang2, J. Shen2, J. Cui2, K. Cai3 and B. G. Thomas4

Optical microscopy, total oxygen measurements and slime tests have been conducted to quantify the size distribution and amount of inclusions at various processing steps during basic oxygen furnace (BOF) based production at Baosteel. The effects on steel cleanliness of specific operational improvements during steel refining and continuous casting have been investigated. Such improvements to these processes and the resulting level of steel cleanliness at Baosteel are summarised in the present paper. Ladle reduction lowers FeOzMnO in the slag to below 5%, decreasing steel reoxidation by the slag. treatment by CaSi wire injection during ladle furnace (LF) refining is used to modify inclusions and improve submerged entry nozzle (SEN) clogging. Slag detection is employed at the ladle bottom during continuous casting. Flow control devices, CaO containing filters and high CaO based basic powder with CaO/SiO2.4 are used in the tundish to remove more inclusions. Several improvements to the castability and in the attainment of clean steel at mould operations have also been made. With this BOF based steelmaking process, impurity levels can be controlled to achieve total oxygen (TO),16 ppm, [S],5 ppm, [P],35 ppm, [N],29 ppm, [H],1 ppm in line pipe , and [C],16 ppm, TO,19 ppm, [N],15 ppm in interstitial free (IF) steels.

Keywords: Clean steel, Inclusions, Impurity elements, Interstitial free steel, Line pipe steel

Introduction low , low and suitable Ca/S; hydrogen induced cracking (HIC) resistant steel requires The importance of clean steel in terms of product quality P(50 ppm and S(10 ppm; and bearing steel requires is increasingly being recognised. Clean steel requires the total oxygen to be less than 10 ppm.3 In addition, control of the size distribution, morphology and many applications restrict the maximum size of inclu- composition of non-metallic oxide inclusions in addition sions,3,4 so the size distribution of inclusions is also to the amount. Furthermore, sulphur, , important. The control of steel cleanliness has been 1,2 hydrogen, nitrogen and even should also be extensively reviewed by Kiessling in 1980 (Ref. 5), controlled to improve the steel properties. For example, McPherson and McLean in 1992 (Ref. 6), Mu and formability, ductility and fatigue strength worsen with Holappa in 1993 (Ref. 7), Cramb in 1999 (Ref. 4) and increasing sulphide and oxide inclusion contents. Zhang and Thomas in 2003 (Ref. 3). Lowering the carbon and nitrogen enhances strain aging Baoshan & Steel Co., Ltd (Baosteel) is currently and increases ductility and toughness. Hardenability and the largest steel company in China. Its annual steel resistance to temper embrittlement can be enhanced by production was 11.5 million tonnes in 2003, 11.9 million reducing phosphorus.1 The definition of ‘clean steel’ tonnes in 2004 and 14.0 million tonnes in 2005. With varies with the steel grade and its end use. For example, regard to the basic oxygen furnace (BOF) based interstitial free (IF) steel requires both carbon and steelmaking route, there are three 300 t and two 250 t nitrogen to be ,30 ppm; line pipe steel requires sulphur, BOFs; several steel refining units, including one CAS- phosphorus and total oxygen (TO) all to be ,30 ppm, OB unit (controlled argon stirring–oxygen blow), two RH (Ruhrstahl–Heraeus) degassers and one ladle furnace (LF); and two 1930 mm width slab casters, 1Department of Materials Science and Engineering, Norwegian University two 1450 mm width slab casters and one 2300 mm width of Science and Technology (NTNU), Alfred Getz vei 2, N–7491 heavy plate caster. Since 1990, efforts to improve steel Trondheim, Norway 2Baosteel Co., Shanghai 201900, China cleanliness have focused on developing steelmaking 3School of Metallurgical Engineering, University of Science and practices to lower TO, N, S, P, H and C levels to Technology Beijing, Beijng 100083, China 4Department of Mechanical and Industrial Engineering, University of achieve low carbon aluminium killed (LCAK) steel, Illinois at Urbana-Champaign, 1206 West Green St, Urbana, IL61801, ultra LCAK steel, such as IF steel, and line pipe steel, as USA well as further improve the castability by using cleaner *Corresponding author, email [email protected] steel or by using special methods such as slag detection

ß 2006 Institute of Materials, Minerals and Mining Published by Maney on behalf of the Institute Received 8 July 2005; accepted 23 November 2005 DOI 10.1179/174328106X80127 Ironmaking and Steelmaking 2006 VOL 33 NO 2 129 Zhang et al. BOF based steelmaking and cleanliness control at Baosteel

1 Sampling locations for continuously cast slab: TO total oxygen

during pouring, techniques to prevent nozzle clogging, extract the inclusions using the slime test,3 which is a more resistant linings and breakout pre- method of extracting inclusions from steel via electro- diction systems at the caster. For LCAK steel and IF lysis, leaving the basic components such as CaO, MnO, steel, the production process is BOFRRHRcontinuous MgO, etc. In the case of slab samples, as shown in casting (CC), and for line pipe steel, the process is Fig. 1, 20620 mm square section bars were taken BOFRRHRLFRCC. The present paper describes through the whole thickness (250 mm), and were cut industrial clean steel production using the BOF based into 10 small pieces for microscope examination, TO steelmaking process and steel cleanliness control at and nitrogen measurement. Finally, 706706150 mm Baosteel. steel samples were cut from the slab for slime extraction, after the scale and the first few millimetres of the surface Experimental method and examination were machined off. Analysis included the chemical composition of slag and steel samples, microscope of inclusions in steel examination of microinclusions, slime extraction of Experimental methods macroinclusions and SEM analysis of the morphology and composition of inclusions. Inclusions on the section In the present work, several slag and molten steel surface of ,300 steel samples from the ladle, tundish samples were taken before, during and after steel and slab were also investigated. On each sample, 300 refining, from the ladle, tundish and mould, and samples random areas with a diameter of 0.3 mm were examined were removed from various locations in the slab. Ladle under the microscope to give the number and size of steel samples were taken 500–600 mm below the top slag inclusions; thus, 21.2mm2 total area was observed for in the ladle, tundish steel samples from 300 mm above each sample. its outlet and mould steel samples from 150 mm below In the present work, ‘macroinclusions’ were those the meniscus and 300 mm away from the submerged greater than 50 mm in diameter. Most of these were entry nozzle (SEN) outports. The sampler was a detected in the residues extracted by electrolytic isola- cylindrical steel cup with a cone shaped cover tion (‘slime test’) from the larger steel samples. The to protect it from slag entrainment during immersion. ‘microinclusions’ data derive from microscopic assess- Attached to a long bar, the sampler was immersed deep ments carried out on planar sections, most of which into the molten steel, where the copper melted and the were smaller than ,50 mm. cup was filled. Small steel samples from the ladle, tundish and mould, 80 mm in length and 30 mm in diameter, were machined into 5 (dia.)65 mm cylinders Morphology and composition of typical for TO and nitrogen analysis, and 20 (dia.)615 mm inclusions cylinders for microscope examination. The steel powders The morphology, composition and likely sources of resulting from machining were used for analysis of the typical inclusions found in LCAK steel samples from the carbon, phosphorus and sulphur contents. Large steel ladle, tundish and mould are shown in Figs. 2 and 3 and samples from the ladle and tundish, 200 mm in length given in Tables 1 and 2, respectively. The morphologies and 80 mm in diameter, were machined into 60 included: (a) angular aluminate (Fig. 2d and f and (dia.)6150 mm cylinders; the steel was dissolved to Fig. 3b); (b) alumina cluster (Fig. 2b and e); and (c)

a ladle; b tundish; c, d mould; e, f slab 2 Typical inclusions from given samples examined by microscope (see Table 1)

130 Ironmaking and Steelmaking 2006 VOL 33 NO 2 Zhang et al. BOF based steelmaking and cleanliness control at Baosteel

unit two-dimensional section area according to micro- scope examination. Figure 6 illustrates the weight of large inclusions per 10 kg of steel extracted by the slime method, which is similar to the inclusion mass fraction in the steel. Inclusions extracted by slime tests that were smaller than 62 mm were suspended in water and their size distributions measured with a Coulter counter, to obtain a three-dimensional inclusion size distribution. The curves were extrapolated to ,120 mm, as shown in Fig. 7, by comparing the measured amount of extracted inclusions larger than 50 mm. The inclusion mass fraction was 66.8 ppm in the tundish, 57.7 ppm in the 20 mm thickness nearest the slab surface and an average of 51.9 ppm in the slab. This suggests that the inclusion content in the interior of the slab (i.e. except the outer 20 mm thickness of the slab) was 50.6 ppm. The fraction of inclusions removed from the tundish to the slab was ,22%. Figure 7 also indicates that the number density (m–3 of steel) of inclusions decreased with increasing size; nevertheless, in terms of volume (or mass) fraction, 10–40 mm inclusions constituted 68–79% of the total fraction of inclusions in the steel, and 40–60 mm inclusions constituted 9–19%. The inclusion size distribution according to two- dimensional (2D) microscope examination (Fig. 5) can be converted into a three-dimensional (3D) distribution a tundish; b slab using 3 Typical inclusions from given samples extracted using n ~ 2D | 12 slime method (see Table 2) n3D 10 (1) dp –2 spherical silicate (Fig. 2a and c and Fig. 3a). The where n2D is the number of inclusions mm of steel surface area, dp is the inclusion diameter in mm and n3D possible sources were deoxidation products, reoxidation –3 products, slag entrapment or broken refractory lining is the number of inclusions m of steel volume. This bricks. Inclusions larger than 50 mm in LCAK steel and expression assumes that each inclusion is approximately line pipe steel were now less than 1 mg/10 kg steel, cylindrical in shape, with height (into the plane) equal which corresponds to ,0.1 ppm TO. These large to its observed diameter. The 3D inclusion size distribu- inclusions could induce serious quality problems owing tion in the tundish and slab converted from 2D examina- to their size, even though their fraction was very small. tion by microscope is show in Fig. 8. By assuming –3 In line pipe steel, besides these common inclusions, 3000 kg m inclusion density, the total mass fraction of . many nanoscale inclusions were found along grain these inclusions can be estimated to be 68 8 ppm in the . boundaries. These nano TiN inclusions changed from tundish and 66 5 ppm in slab. This estimated inclu- square to ellipsoid if combined with Ti O , as shown in sion mass fraction in the tundish is very close to the 2 3 . Fig. 4.8 measured value of 66 8 ppm. Although the estimated inclusion mass fraction in the slab is larger than the Amount and size distribution of inclusions measured value of 51.9 ppm, equation (1) can be used Three kinds of inclusion size distributions are shown in reasonably to convert the 2D size distribution into a 3D Figs. 5–7. Figure 5 illustrates the inclusion number per distribution.

Table 1 Typical inclusions in steel examined by microscope, corresponding to Fig. 2, wt-%

Inclusion in Fig. 2 Al2O3 SiO2 CaO MnO FeO MgO K2ONa2O TiN Possible source a 1.256.40.55.820.310.10.35.10.3 Ladle slag b 71.80.7 ??? 0.224.3 ??? ??? ??? 2.9 Deoxidation or reoxidation product c 13.744.028.70.82.52.90.13.73.7 Mould d 98.40.4 ??? ??? ??? ??? ??? 0.1 ??? Deoxidation product e 94.43.4 ??? ??? 1.8 ??? ??? ??? 0.4 Deoxidation or reoxidation product f 92.82.7 ??? ??? 2.9 ??? ??? ??? 1.6 Refractory

Table 2 Typical inclusions in steel extracted using slime method, corresponding to Fig. 3, wt-%

Inclusion in Fig. 3 Al2O3 SiO2 CaO MnO FeO MgO K2ONa2O TiN Possible source a,1 2.783.2 ??? 9.13.71.1 ??? ??? 0.2 Slag 65.120.49.0 ??? 4.51.0 ??? ??? ??? Slag a,3 24.029.04.119.716.41.40.71.3 ??? Slag b,4 75.419.4 ??? 1.02.02.3 ??? ??? ??? Deoxidation product

Ironmaking and Steelmaking 2006 VOL 33 NO 2 131 Zhang et al. BOF based steelmaking and cleanliness control at Baosteel

a

b

a compound inclusions with composition Ti2O3zMnS; b TiN inclusion 4 Nanoprecipitates in line pipe steel

Total oxygen measurement is an indirect method of Thus, the TO content actually represents the level of evaluating oxide inclusions in a steel.3 The total oxygen ,50 mm small oxide inclusions only. The current TO in (TO) in the steel is the sum of the free oxygen (dissolved IF and line pipe steel slabs at Baosteel is ,16 ppm. The oxygen) and the oxygen combined as non-metallic TO in the ladle, tundish, mould and slab in two typical inclusions. Free oxygen, or ‘active’ oxygen, can be sequences of LCAK steel is shown in Fig. 9, indicating measured relatively readily using oxygen sensors. It is that the TO decreased from the ladle to the tundish, to controlled mainly by equilibrium thermodynamics with the mould and to the continuously cast slab. deoxidation elements, such as aluminium. If [%Al]5 0.03–0.06, the free oxygen is 3–5 ppm at 1600uC. Inclusion distribution in slab Because the free oxygen does not vary much, the total The distribution of inclusions through the thickness of oxygen is a reasonable indirect measure of the total the LCAK steel slab (250 mm thickness, 1300 mm amount of oxide inclusions in the steel. Owing to the width) measured by microscope examination is shown small population of large inclusions in a steel and the in Fig. 10. The data suggest that: (a) inclusions small sample size for TO measurement (normally concentrated more in the 20 mm thickness nearest the ,20 g), it is rare to find a large inclusion in a sample. slab surface; (b) inclusions sometimes accumulated at Even if a sample contains a large inclusion, it is probably between one-half and one-quarter of the slab thickness discounted because of the anomalous high reading. from the inner radius; and (c) filters in the tundish were effective at lowering microinclusion levels. Further investigation by sulphur print detection indicated that this inclusion accumulation was more prevalent in the slab head and tail cast during unsteady conditions at the beginning and end of a sequence, as shown in Fig. 11. Microscope examination and SEM detection suggested that this inclusion accumulation was mainly induced by the entrapment of dislodged clogged materials from the SEN during ladle changes. The composition of a typical inclusion in the slab head containing both clogged material and some broken SEN surface refractory is given in Table 3. Examination also indicated that slag inclusions were mainly entrapped 0–50 mm from the surface of the slab.

Ladle operations to remove more inclusions Ladle slag reduction treatment Reoxidation to form alumina in the ladle during steel refining is mainly caused by SiO2 in the slag and lining 5 Microinclusion size distribution according to micro- refractory, and MnO and FeO in the ladle slag, by the scope examination: RH Ruhrstahl–Heraeus following reactions

132 Ironmaking and Steelmaking 2006 VOL 33 NO 2 Zhang et al. BOF based steelmaking and cleanliness control at Baosteel

7 Inclusion size distribution evolution according to Coulter counter measurement of slime extracted a one weir containing CaO filters and not touching bot- tom of tundish, two dams each side; b same as a inclusions except weir touching bottom Slag reduction treatment is carried out by adding 6 Size distributions of .50 mm inclusions according to aluminium and lime onto the top of the ladle slag to slime extraction using given tundish flow control devices reduce its FeO and MnO content. The effect of ladle slag reduction treatment on the TO content in the steel is 3=2(SiO )z2 Al ~(Al O )z3=2 Si (2) 2 ½ 2 3 ½ shown in Fig. 12. A larger FeOzMnO content in the ladle slag corresponds to higher total oxygen. With the z ~ z 3(MnO) 2½Als (Al2O3) 3½Mn (3) slag reduction treatment, MnO and FeO in the ladle slag were reduced to ,5%, corresponding to ,20 ppm TO in z ~ z 3(FeO) 2½Als (Al2O3) 3½Fe (4) the tundish.

Table 3 Composition of submerged entry nozzle (SEN) clogging materials and typical inclusion accumulated at quarter thickness of slab head, wt-%

Al2O3 SiO2 FeO Na2OCr2O3 CaO ZrO2 SFC

Tundish powder 1.578.91.1 ??? ??? ??? ??? ??? ??? 10.5 Mould flux 1.739.4 ??? 12.8 ??? 36.9 ??? ??? 5.53.5 Initial layer of SEN 40.34 37.19 19.26 0.57 1.09 0.37 0.49 0.69 ??? ??? Clogging materials 1 97.47 2.37 ??? 0.08 ??? ??? ??? ??? ??? ??? Clogging materials 2 72.82 24.31 0.78 0.42 0.83 0.84 ??? ??? ??? ??? Clogging materials 3 92.26 3.65 3.54 0.16 ??? 0.62 ??? 0.03 ??? ??? Typical inclusion 90.93 2.24 3.92 ??? ??? 0.38 1.78 0.74 ??? ???

Ironmaking and Steelmaking 2006 VOL 33 NO 2 133 Zhang et al. BOF based steelmaking and cleanliness control at Baosteel

11 Inclusion distribution along slab thickness according to sulphur print

RH treatment Figure 13 shows that RH treatment removed up to 90% of macroinclusions, 50–70% of microinclusions and 70– 90% of the TO. Alumina in the ladle slag during RH 8 Three-dimensional size distribution of inclusions con- treatment increased 8–13% from its original content. verted from two-dimensional microscope examination The results for different RH deoxidation times suggest an optimum treatment time of 8–12 min (Fig. 14). Beyond this optimum time, the oxygen removal effi- ciency may decrease owing to refractory lining erosion. Also, excessive stirring is detrimental because it may expose an ‘eye’ or slag free region of the steel surface to air reoxidation and perhaps even slag entrainment. Calcium treatment Nozzle clogging induces serious castability problems in aluminium killed steels, such as lowering the casting speed, inducing asymmetrical fluid flow and level fluctuations in the mould, thus entrapping more inclu- sions, and sometimes causing a breakout. Removing more inclusions before continuous casting is the best way to prevent nozzle clogging, and is the only approach for steels with very strict requirements on formability, because calcium treatment at the ladle may generate new hard inclusions.9 For other aluminium killed steels, and 9 Total oxygen in steel from ladle to slab

10 Microinclusion distribution along slab thickness with (strand 3) and without (strand 4) tundish filters: one 12 Effect of FeO and MnO content in ladle slag on TO in weir and one dam each side in tundish steel

134 Ironmaking and Steelmaking 2006 VOL 33 NO 2 Zhang et al. BOF based steelmaking and cleanliness control at Baosteel

15 Submerged entry nozzle (SEN) clogging in heat with 13 Steel cleanliness before and after RH treatment: large calcium treatment during ladle refining index means poor cleanliness of mainly CA2. Current practice at Baosteel indicates that [Ca] should be .25 ppm in order to prevent solid for greatly desulphurised line pipe steels, calcium alumina based inclusion clogs (Fig. 16). Too much treatment can be used to improve SEN clogging. At calcium can also generate CaS with a high melting point Baosteel, CaSi wire is fed into the molten steel during (2450uC). Increasing dissolved aluminium also decreases ladle refining. Alumina reacts with CaO, forming the oxygen activity, generating sulphide inclusions. Too calcium aluminates. If the generated calcium aluminates much sulphur in the steel and too low a temperature also have a low melting point, then clogging is improved. The enables CaS generation. Baosteel practice indicates that possible compound inclusions generated by calcium ,50 ppm [Ca] in the steel can prevent CaS generation, . treatment include CA6,CA2, CA, C12A7 and C3A, where and [Ca]/[Al].0 09 favours prevention of nozzle clog- C and A represent CaO and Al2O3, respectively. The ging (Fig. 17). Hence, [Ca] needs to be controlled within first two should be avoided owing to their high melting the range 25–50 ppm, and [Ca]/[Al].0.09, to avoid point over 1700uC. Adding too much or too little nozzle clogging problems. calcium can also induce SEN clogging, however. An example of SEN clogging during continuous casting of a Ladle slag detection during pouring calcium treated heat with 16 ppm [Ca] after calcium An electromagnetic slag detection system10 is used at the treatment is shown in Fig. 15. The white centre is steel, bottom of the ladle during pouring to minimise slag and the black surround is clogging, with a composition carryover at the end of pouring, induced by the swirl flow. Slag inclusions in the steel are decreased and consequently the steel yield is improved, as shown in Fig. 18. The steel left in the ladle is 1 t less than before.

14 Effect of RH degassing time on TO in tundish 16 Effect of calcium content in steel on nozzle clogging

Ironmaking and Steelmaking 2006 VOL 33 NO 2 135 Zhang et al. BOF based steelmaking and cleanliness control at Baosteel

18 Improvement of casting yield using slag carryover detection system at ladle 17 Effect of Ca/Al ratio on nozzle clogging Basic tundish flux

Tundish and caster Plant experience found that MgO–SiO2–Al2O3 based operations to improve steel cleanliness basic tundish flux tended to freeze owing to its high melting point (.1400uC). This made it difficult to Flow control devices absorb inclusions, and also could not effectively prevent The two strand tundish investigated has a steady air absorption. Instead, CaO–SiO2–Al2O3 based fluxes . operating capacity of 60 t, 1 2 m depth and MgO based were developed. The effect of the CaO/SiO2 ratio on basic refractory lining. The effect of dams and filters on castability and steel cleanliness has been investigated. inclusion removal has been investigated. Table 4 indi- The main characteristics of the fluxes are given in cates the removal of TO, microinclusions and macro- Table 5. The molten steel temperature and TO in the inclusions between the ladle after RH treatment and the tundish during IF steel casting are shown in Fig. 19. The tundish. More inclusions were removed in the tundish steel temperature during a ladle exchange period was with each side having two dams, and a weir containing more steady and lower TO was achieved when the flux filters, than in the tundish with only one dam, and a weir contained CaO/SiO2.4, compared with CaO/SiO2,2. without filters. The filters were effective at removing In addition, less erosion of the tundish and SEN more TO and macroinclusions. There was no clear refractory lining was found for the tundish flux with difference in steel cleanliness whether the weir touched CaO/SiO2.4 than with CaO/SiO2,2. the bottom or not (Fig. 6). Caster operations Many techniques are used at the caster to produce ultraclean steel, including: Table 4 Measured inclusion content and removal (i) low carbon, high mould flux for fractions from ladle to tundish* LCAK steel Heat 1 Heat 2 Heat 3 (ii) low sulphur and phosphorus content mould flux for line pipe steel Total oxygen (TO), ppm (iii) mould level fluctuations controlled within Ladle 103 70 67 ¡ Tundis 81 63 44 3mm Removal fraction 21% 10% 34% (iv) argon injection optimised effectively to protect Microinclusions, mm–2 molten steel from air absorption, and to Ladle 7.28.410.8 maintain a stable double roll flow pattern in . . . . Tundish 5 16032–5 1 the mould Removal fraction 29% 28% 53–70% (v) SEN refractory lining chosen to prevent nozzle Macroinclusions, mg/10 kg steel Ladle 138.966.094.3 clogging and retard erosion by mould flux Tundish 36.932.69.4 (vi) effective breakout prediction system implemen- Removal fraction 74% 51% 90% ted using three rows of at the mould copper plates. *Heat 1: one weir (no filters) and one dam each side; heat 2: one weir (containing CaO filters) and one dam each side; heat 3: one Since Baosteel developed its own breakout prediction weir (containing CaO filters) and two dams each side (weir did system in 1999, all sticker breakouts have been predicted not touch tundish bottom). and avoided during the continuous casting of 3263

Table 5 Parameters of CaO based basic tundish flux

CaO/SiO2 Al2O3, % MgO, % F, % Melting point, uC Viscosity at 1400uC, Pa s

4.0–8.0 30–50 5.0–10.0 ,5.0 1200–1400 0.20–1.0 0.8–2.0 ,10.0 ,15.0 ,5.0 1190–1350 0.20–1.0

136 Ironmaking and Steelmaking 2006 VOL 33 NO 2 Zhang et al. BOF based steelmaking and cleanliness control at Baosteel

19 Temperature and total oxygen (ppm) of molten steel in tundish (four heats) during continuous casting 20 Distributions of total oxygen, sulphur, phosphorus using given tundish fluxes and carbon along slab thickness heats. The false alarm rate is nine per 10 000 heats. This critical to prevent air absorption, which is the source of system has enabled a higher average casting speed, and nitrogen pickup. Oxygen pickup is always many times fewer defects owing to cracks and slag entrainment greater than the measured nitrogen pickup, owing to its during transient casting. Currently there are almost no faster absorption kinetics at the air/steel interface.11 In sticker breakouts, only ,5 breakouts per year from addition, nitrogen pickup is faster when the oxygen and longitudinal cracks. sulphur contents are low.12 Thus, to reduce nitrogen pickup, deoxidation is best carried out after tapping, Control of nitrogen, carbon, sulphur and which is the current practice for clean steel grades at phosphorus in steel Baosteel. The current nitrogen content of IF steel and Nitrogen line pipe steel slabs is 15–30 ppm, and nitrogen pickup from ladle to tundish can be controlled below 1.5 ppm Normally, a large nitrogen content at tapping tends to by optimised shrouding, argon gas injection and fibre result in a large nitrogen content in the slab. Thus, the sealing at the tundish and SEN. control of nitrogen should mainly focus on lowering the nitrogen content during BOF blowing and preventing nitrogen pickup during tapping, steel refining and Carbon continuous casting. Currently, at Baosteel, nitrogen The greatest decarburisation is for IF steel by RH during BOF steelmaking fluctuates from 11 to 43 ppm. treatment. Techniques to improve this operation include: Plant experiments indicate that when [N] is less than (i) optimising initial [C] and [O] before degassing 30 ppm before RH treatment, [N] cannot be lowered into the ranges of 500–650 ppm and 300– further by RH treatment. The shrouding system is 450 ppm, respectively

Ironmaking and Steelmaking 2006 VOL 33 NO 2 137 Zhang et al. BOF based steelmaking and cleanliness control at Baosteel

(ii) enlarging the snorkel diameter from 500 to (i) de-Si, de-P and de-S at hot treatment, 750 mm and increasing the argon flowrate from followed by BOF steelmaking with a small slag 1000 to 3000 NL min–1. content, lowering [P] to 120 ppm After RH treatment, [C] can be lowered to 10 ppm. The (ii) de-S at hot metal treatment, then the BOF [C] pickup occurring during continuous casting is process with a large slag content, lowering [P] to controlled below 6 ppm by the following techniques: 100 ppm (i) using low carbon, high viscosity mould flux, (iii) de-Si, de-P and de-S at hot metal treatment, decreasing carbon pickup at the continuous followed by BOF steelmaking with a large slag casting mould from 5.5to1.8 ppm content (slag content index 1.0), lowering [P] to (ii) using carbon free ladle refractory lining 66 ppm (iii) using high basic, low carbon tundish flux (CaO/ (iv) de-P at hot metal treatment, followed by BOF SiO2.4) steelmaking with a large slag content (slag (iv) maintaining thick liquid flux layer and control- content index 0.6), lowering [P] to 58 ppm ling mould fluid flow to lower the standing (v) double BOF steelmaking process, achieving wave and level fluctuation. 20 ppm [P] in the steel. Sulphur The control of impurity elements at Baosteel has improved considerably during the past 15 years, as The initial sulphur content of the molten iron at indicated in Table 6. Baosteel steel can now achieve Baosteel is ,200 ppm. After hot metal desulphurisation TO,16 ppm, [S],5ppm, [P],35 ppm, [N],29 ppm, by injection of CaC2 powder or based [H],1 ppm in line pipe steel, and [C],16 ppm, powder, the sulphur decreases to 10–30 ppm. It is TO,19 ppm, [N],15 ppm in IF steel. Currently, important to remove the top slag quickly after [S]z[P]zTOz[N]z[H] in line pipe steel can be desulphurisation in order to decrease sulphur pickup. maintained below 85.5 ppm, and [C]zTOz[N] in IF During the BOF steelmaking process, there is 10– steel can be kept below 50 ppm. Figure 20 shows 30 ppm sulphur pickup, mainly from lime and . example distributions of TO, sulphur, phosphorus and To achieve an ultralow sulphur content, especially for carbon along the thickness of an ultraclean steel slab line pipe steels, three desulphurisation methods during produced in 1997. There is a slight increase in impurities the steel refining process have been developed: towards the inside radius, owing to inclusion flotation (i) CaO–CaF2 flux is added to the vacuum chamber and a very slight centreline segregation. through the addition hoppers during RH treatment; slag carryover from the BOF is controlled carefully for these heats, and ladle slag Conclusions reduction treatment is carried out to decrease the 1. Techniques to improve steel cleanliness at Baosteel FeO content in the slag before steel desulphurisa- include ladle slag reduction treatment to lower FeO and tion; [S] is lowered from 28.4to16.2ppm MnO in the ladle slag to below 5% before steel refining, (ii) during RH treatment, after strong deoxidation, suitable CaSi wire injection in the ladle, ladle slag CaO–CaF2 powder is injected into the molten detection during pouring, development of a CaO based steel in the ladle by a lance below the up snorkel; basic tundish flux, optimisation of flow control devices [S] can be lowered from 61.9to35.8 ppm in the tundish and optimisation of mould flow to avoid (iii) desulphurisation can be achieved at the LF by mould slag entrainment. optimising argon blowing to obtain a suitable 2. Inclusions concentrate mostly within 20 mm of the emulsification condition, improving the reaction slab surface. Some slabs experience occasional accumula- between slag and molten steel; [S] can be lowered tion at one-quarter to one-half slab thickness from the from 67.0to8.7 ppm. This method can lower the inner radius, mainly induced by the entrapment of released sulphur below 10 ppm, so is currently used for clogged materials from the SEN during ladle exchanges. the production of ultraclean line pipe steel. It 3. Castability has been improved by an improvement should be noted that the MgO–CaO refractory of steel cleanliness, the use of optimal calcium treatment lining and tundish flux may also remove some to prevent nozzle clogging and application of a breakout sulphur by the following reaction prediction system at the caster. 4. Currently, the impurity elements in steel can be (CaO)z2=3½Alz½S~(CaS)z1=3½Al O (5) 2 3 controlled to TO,16 ppm, [S],5 ppm, [P],35 ppm, [N],29 ppm, [H],1 ppm for line pipe steels, and Phosphorus [C],16 ppm, TO,19 ppm, [N],15 ppm for IF steels. Five different processing routes are used to achieve low phosphorus steel at Baosteel: References Table 6 Impurity content of line pipe steel and inter- 1. K. W. Lange: Int. Mater. Rev., 1988, 33, 53–89. stitial free (IF) steel achieved at Baosteel, ppm 2. W. B. Morrison: Ironmaking Steelmakaing, 1989, 16, 123–130. 3. L. Zhang and B. G. Thomas: ISIJ Int., 2003, 43, 271–291. Line pipe steel IF steel 4. A. W. Cramb: in ‘Impurities in engineered materials: impact, reliability and control’, (ed. C. L. Briant), 49–89; 1999, New York, Year [S] [P] TO [N] [H] [C] [N] TO Marcel Dekker. 5. R. Kiessling: Met. Sci., 1980, 15, 161–172. 1996 32 134 35 47 2 50 24 50 6. N. A. McPherson and A. McLean: ‘Continuous casting’, Vol. 7, 1999 16 89 24 30 2 23 16 28 ‘Tundish to mould transfer operations’; 1992, Warrendale, PA, ISS. 7. D. Mu and L. Holappa: ‘Production of clean steel: literature 2003 9 54 16 30 1.5161519 survey’, Report no. PB93–179471/XAB, Gov. Rep. Announce. 2004 4.835.015.829.01.0 ??? ??? ??? Index (USA), 1993, 44.

138 Ironmaking and Steelmaking 2006 VOL 33 NO 2 Zhang et al. BOF based steelmaking and cleanliness control at Baosteel

8. Q. Zheng, Z. Chen and L. Zhu: Proc. 13th CSM Annu. 10. P. Andrzejewski and K.-U. Kohler: ISS Steelmaking Conf. Proc., Steelmaking Conf., Kunming, China, 2004, Chinese Society for 1990, 73, 41–42. . 11. K. Sasai and Y. Mizukami: ISIJ Int., 2000, 40, 40–47. 9. K. Larsen and R. J. Fruehan: Iron Steelmaker (ISS Trans.), 1991, 12. N. Bannenberg and K. Harste: Re´v. Me´tall., Cah. Inf. Tech., 1993, 12, 125–132. 90, 71–76.

Ironmaking and Steelmaking 2006 VOL 33 NO 2 139